An evolutionary model for sleep quality analytics using fuzzy system.

Area of Science:

• Neuroscience
• Computer Science
• Artificial Intelligence

Background:

• Electroencephalography (EEG) signals reflect brain activity crucial for determining sleep stages.
• Manual sleep stage classification is time-consuming and subjective.
• Automated methods using machine learning offer objective and efficient alternatives.

Purpose of the Study:

• To develop and evaluate an automated method for sleep stage classification using EEG signals.
• To compare the performance of a fuzzy min-max neural network against various machine learning and deep learning models.
• To assess the effectiveness of extracted EEG signal patterns for sleep stage identification.

Main Methods:

• Implementation of a fuzzy min-max neural network for sleep stage classification and clustering.
• Comparison with established algorithms: KNN, Random Forest, Decision Tree, XGBoost, AdaBoost, LDA, QDA, and CNN.
• Feature extraction and pattern analysis from EEG signals for model training.

Main Results:

• The fuzzy min-max classifier achieved the highest accuracy at 86%.
• Convolutional Neural Network (CNN) followed with 81% accuracy.
• Other machine learning models showed significantly lower accuracies, with Random Forest at 55.46%.

Conclusions:

• The fuzzy min-max neural network demonstrates superior performance for automated sleep stage classification.
• The study highlights the potential of fuzzy logic and deep learning (CNN) in advancing sleep analysis.
• Accurate automated sleep stage analysis is feasible and beneficial for research and clinical applications.